We propose a method for answering critical questions concerning the social determinants of health and health disparities by linking 30 years of comprehensive, nationally-representative sociological data to prospective mortality data containing specific causes of death. The problem: African-Americans and persons of low socio-economic status (SES) tend to have less access to social resources, such as good schools, than do whites and persons of high SES, respectively. Over time, disparities in access to such social resources translate into health disparities. For instance, disparities in access to quality education between groups results in occupational disparities that lead to income disparities. Those with less education and income have not only lower access to health insurance, but also less access to banking services, transportation, and modes of democratic engagement, as well as a host of other social systems. Researchers and thinkers dating back to Hippocrates have observed that such social deprivations lead to health disparities, but it was long thought that health disparities primarily arose from lower access to lifesaving material goods. In the twentieth century, researchers turned their attention to the association between social deprivations and psychosocial possesses as additional causes of health disparities. For instance, lower social capital (e.g., trust, group participation), social ties (e.g., friends and family relationships), and harmful psychological states (e.g., stress, pessimism) have been identified as contributors to health disparities by race, ethnicity, and SES. Despite the importance of understanding the root social causes of health disparities, however, a prospective, comprehensive sociomedical dataset has never been developed. Our proposed solution. Scientists do have access to datasets that allow for the description of health disparities. These datasets consist of a number of important medical risk factors and outcomes alongside information about the subjects'income and educational attainment. However, the datasets do not permit a deeper analysis of social and psychological causes of these disparities. Such a dataset would not only need to contain more comprehensive questions about the subjects'childhood and adult SES, they would also require questions about subjects'social networks, thought, feelings, attitudes, beliefs, and participation in civil society. In short, scientists require a dataset that allows exploration of putative social and psychological causes of health disparities. Clearly, development of an interdisciplinary, nationally-representative, prospective dataset containing robust measures in social and psychological domains would cost millions of dollars and would require many years of follow-up. Our proposal provides a rapid and highly cost-effective shortcut that will produce an extremely robust dataset with outstanding follow-up and oversampling of minority populations. Specifically, we propose to link the 1977-2007 General Social Survey (GSS), a multiple-year, cross-sectional survey rich in health and sociological variables, to National Death Index (NDI) data through 2008. This dataset will advance social epidemiologic studies of health disparities beyond mere identification and description to a deeper understanding of the underlying mechanisms. This will usher in highly targeted policies to address "health gaps" between groups. To catalyze the utility and the widespread use of this dataset, we will 1) release the dataset to the wider research community, 2) convene a group of leading transdisciplinary experts in order to troubleshoot and disseminate the GSS-NDI data, and 3) develop a set of useful research tools that will facilitate the adoption and reach of the GSS-NDI. These tools include a means for defining multi-dimensional concepts such as social capital, a means for adjusting for between group differences in the interpretation of qualitative questions (i.e., African Americans may be more likely than whites to see the question, "People treat me fairly," as having a racial component), and, finally, a means for translating mortality differences into life expectancy differences so that the policy implications of researchers'findings can be more easily understood. In short, our project will create the prospective sociomedical dataset that has been missing in the health sciences, a dataset capable of greatly advancing our understanding of the non-medical determinants of health and their relationship to health disparities. We will create this dataset in a fraction of the time and at a fraction of the cost of generating a prospective sociomedical dataset from scratch. Finally, we have done preliminary work to both ensure its success and to ensure that our project is ready for rollout. We forward this proposal under the RC2 (GO) mechanism to the National Center on Minority Health and Health Disparities'Social Determinants of Health Initiative. We propose to link over thirty years of data from the longest running social science dataset to mortality data, creating the first major prospective sociomedical dataset. This dataset will greatly advance scientific understanding of the social causation of disease and health disparities.